Commentary/Erdelyi: The unified theory of repression
recollections” (Freud 1901/1960, p. 6), relates specifically to his
theory of the forgetting of proper names. The quoted words
concern a rather far-fetched analysis he has just given of the forgetting of a person’s name and its (tentative) replacement by a
different one. However, the “mechanism of false recollection”
as described here requires a process of repression, not one of
modification over time. Moreover, Timpanaro has examined
this very example that Freud uses to exemplify his theory of
the forgetting of names, and shown that it can be explained perfectly adequately without appealing to Freudian repression processes (Timpanaro 1976, pp. 63 – 65, 70 – 71, 75 – 77, 85 – 89).
Although Erdelyi is correct in noting the similarity between
reconstructions based on emotional motives and those based on
cognitive schemes, neither of these incremental reconstructive
biases resemble the production of erroneous memories as
described in the passages he cites. Of course it is possible to highlight certain elements in Freud’s writings about memory that
overlap with Bartlettian theories (e.g., his saying that “recall is an
amalgam of fact and fiction”), but if the full context of Freud’s
accounts are taken into consideration, it is apparent that every
one of the truncated quotations that Erdelyi cites in the paragraph
in question relates to theories of memory that bear little relationship to Bartlettian reconstructive processes over time.
The social psychology of cognitive repression
Jennifer J. Freyd
Department of Psychology, 1227 University of Oregon, Eugene, OR
97403 – 1227.
jjf@dynamic.uoregon.edu
http://dynamic.uoregon.edu/
Abstract: Erdelyi identifies cognitive and emotional motives for repression,
but largely neglects social motivations. Yet social pressure to not know, and
implicit needs to isolate awareness in order to protect relationships, are
common motives. Social motives may even trump emotional motives; the
most painful events are sometimes the most difficult to repress. Cognitive
repression may be impacted by social information sharing.
That people sometimes distort, omit, deny – repress 1 – information is indeed obvious, as articulated in Erdelyi’s insightful and
timely target article. But why do individuals repress information?
Erdelyi explicitly refers to only two categories of motive: cognitive
and emotional. Erdelyi has neglected a third important category of
motive for repression: social. Social motives include all the reasons
individuals repress as a consequence of social pressures to not
know and to doubt memories, and in response to an implicit
need to isolate information from awareness in order to maintain
a whole host of social (intimate, communal, political) relationships.
Campbell (2003) notes that “political struggles over memory
are extraordinarily intense” (p. 52). Whether a struggle over
memory for the events witnessed at Mai Lai (Freyd 2001) or
the more mundane experience of discussing experiences of the
day, social response to testimony will shape what is remembered
(Fivush et al. 1997). Sometimes forgetting and unawareness are
explicitly encouraged by others – for instance, when harmful
events are perpetrated by one person or group of people on
another with manipulations to reduce the subsequent disclosure
and memory of those events (Veldhuis & Freyd 1997).
Individuals are sometimes exposed to information that may
endanger their well-being. In such cases, repression may be
adaptive. Childhood abuse perpetrated by a caregiver is an
example. Betrayal trauma theory (BTT; Freyd 1994; 1996;
2001; see Fig. 1) proposes that the way in which events are processed and remembered will be related to the degree to which a
negative event represents a betrayal by a trusted, needed other
(Sivers et al. 2002). Full awareness of such abuse may only
increase the victim’s risk by motivating withdrawal or confrontation with the perpetrator, thus risking a relationship vital to the
victim’s survival.
518
BEHAVIORAL AND BRAIN SCIENCES (2006) 29:5
Figure 1 (Freyd). Two-Dimensional Model for Traumatic Events.
Social betrayal is hypothesized to be associated with greater
unawareness and reduced recall.
BTT posits that knowledge isolation is predicted by the threat
the information poses to the individual’s attachment system.
Studies indicate reduced recall in the case of abuse by caregivers
or close others (e.g., Edwards et al. 2001; Freyd et al. 2001;
Schultz et al. 2003; Stoler 2001; Sheiman 1999; Williams 1995;
for additional analyses see Freyd 1996). Freyd et al. (2001)
found that physical and sexual abuse perpetrated by a caregiver
was related to higher levels of self-reported less persistent memories of abuse compared to non-caregiver abuse (see Fig. 2).
Exposure to high-betrayal traumas is also associated with
higher levels of dissociative tendencies (Chu & Dill 1990;
DePrince 2005; Plattner et al. 2003). Dissociation, like repression, may help keep betrayal information out of awareness.
DePrince and Freyd (2004) found that under divided-attention,
high dissociators have impaired memory for words associated
with trauma (e.g., incest) but not for neutral words, as compared
with low dissociators. Similarly, Becker-Blease et al. (2004) found
that, under divided attention, abused children remembered
fewer abuse-charged pictures relative to non-abused children.
Figure 2 (Freyd). Average memory impairment (reduced recall
range is 0 for excellent memory to 1.0 for very poor memory for
abuse) for caretaker (high betrayal) and non-caretaker (lower
betrayal) sexual, physical, and emotional abuse. (Results from
Freyd et al. 2001.)
Commentary/Erdelyi: The unified theory of repression
What about betrayals, such as adultery, that do not meet criteria for abuse – are these also repressed? One striking phenomenon is when adultery is obvious to those outside the relationship,
but the wronged partner is apparently unaware of the betrayal. I
have conducted interviews with individuals who later discovered
they had been in exactly this situation. In each case my informants described feeling dependent upon their partners. This
appears to be betrayal blindness (Freyd 1996; 1999) – by not
knowing about the adultery, a crisis is forestalled and a necessary
relationship is “protected.”
Awareness of betrayal is destabilizing. Zurbriggen (2005) has
argued that a similar process can occur when people systematically
remain unaware of observable betrayals by their own government.
One motivation for remaining unaware is likely dependence upon
the government for protection from threat. Shay (1994) also
pointed to betrayal (particularly by commanding officers) as a
key factor in soldiers’ harm from war. War memories are among
those must susceptible to forgetting (Elliott 1997; Grinker &
Spiegel 1945; Sargant & Slater 1941; Thom & Fenton 1920).
It is commonly stated that the primary motive for repression is
avoidance of psychic pain. Yet pain is needed to motivate behavior,
and lack of pain can be dangerous. Rather than avoidance of pain
per se, I argue that repression is most likely when it serves a function such as protecting a necessary relationship. When betrayal is
fully perceived, the resulting pain may motivate withdrawal or confrontation. In a betrayal trauma, this reaction could result in an
even worse situation for the betrayed person; thus repression is
adaptive. There is an analogous effect for physical pain – if mobility is sufficiently urgent for survival, even a serious injury may be
ignored. The usual processes that produce severe pain and related
unwillingness to move can be spontaneously suppressed in the
right circumstances. This is not due to a desire to avoid pain,
but a survival need that overrides the usual pain/immobility
response. So, too, for psychic repression. Consistent with this, it
is apparently more difficult to forget the death of one’s own
child than many other traumatic experiences (Elliott 1997).
Sharing or not sharing information about events has implications for the memorability of those events. A lack of social
sharing of information may lead to decreased opportunities to
rehearse information, encode it in more shareable and stable
forms (Freyd 1983; 1990; 1996), and eventually impact recall
(Freyd 1996; 2003). Consistent with this, child sexual abuse has
a strikingly low disclosure rate and also has some of the highest
rates of reduced recall (Elliott 1997; Freyd et al. 2005). Similarly,
Foynes et al. (2006) found that non-disclosure of childhood abuse
was predicted by the closeness of the perpetrator, just as reduced
recall is so predicted (Freyd et al. 2001).
Although Erdelyi makes great progress toward a unified theory of
repression, he largely neglects a crucial aspect of repression – the
social psychological forces and motives. It is of course the case
that individual nervous systems instantiate psychological repression, but a full understanding of the phenomena will not occur if
the individual psyche is studied in a vacuum. Betrayal and other
social forces must be part of a unified theory of repression.
NOTE
1. The words “repress” and “repression” are contested and often
avoided, but in this context I follow Erdelyi’s target article, which is defining and reclaiming these words.
Forging a link between cognitive and
emotional repression
Esther Fujiwaraa and Marcel Kinsbourneb
a
Rotman Research Institute at Baycrest, Toronto, Ontario, M6A 2E1, Canada;
Department of Psychology, New School University, New York, NY 10003.
efujiwara@rotman-baycrest.on.ca
kinsboum@newschool.edu
b
Abstract: Erdelyi distinguishes between cognitive and emotional forms
of repression, but argues that they use the same general mechanism.
His discussion of experimental memory findings, on the one hand, and
clinical examples, on the other, does indeed indicate considerable
overlap. As an in-between level of evidence, research findings on
emotion in neuroscience, as well as experimental and social/personality
psychology, further support his argument.
Erdelyi’s unifying theory convincingly addresses the extensively
used, yet frequently misunderstood, concept of repression. He
shows the similarities between early definitions in psychodynamic
and experimental psychology, literatures largely oblivious of each
other. His key point is that repression should be understood as a
mechanism regardless of its function, rather than as inherently
defensive. Repression becomes a universal gatekeeper of the contents of consciousness, irrespective of underlying psychodynamic
assumptions and of the nature of to-be-repressed information.
Examples, mostly from the experimental memory literature, illustrate the similarity of distortions that render information inaccessible, whether for cognitive or emotional reasons. The question
remains: If repression as a mechanism comprises suppression,
inhibition, avoidance, and the like, why retain such an embattled
concept? The justification reverts to psychodynamic traditions
and therefore to defensive or emotional forms of repression.
Defensive repression is primarily a phenomenon of emotion,
not memory. Repression of memory content is just one of
many possible consequences. Under some circumstances,
emotional information may be “repressed” on account of an
anticipated or associated unwanted emotional state, rather than
because the memory content per se is unwanted (or instructed
not to be wanted). Erdelyi summons substantial support from
the memory literature for the view that the means by which
these forms of repression function are essentially the same. We
propose to show that, in fact, the link between the experimental
and the clinical findings is even closer than appears from the
evidence presented in the target article.
Erdelyi cautions against inducing unbearable stress in the laboratory so as to mimic trauma and potentially trigger defensive
repression, thereby questioning the validity of inferences from
experimental to clinical findings. However, endocrinological
and social psychology research that utilized various forms of
stress induction found memory performance that is well in line
with Erdelyi’s assumptions. For example, experimental manipulations of stress-hormone levels show evidence for both memory
enhancing and memory-impairing effects. Consistent with
Erdelyi’s comment about temporal shifts from hypermnesia to
amnesia, delicate temporal dynamics of hormonal shifts seem
to be a mediating variable. Administering adrenal glucocorticoid
before memory retrieval generally impairs performance, whereas
before memory encoding, it enhances performance (Het et al.
2005). This dissociation also obtains with psychosocial rather
than direct hormonal stress manipulation (Kuhlmann et al. 2005).
Further, to forge the link between cognitive and emotional forms
of repression, the extensive body of literature on the so-called
repressive coping style is worth considering in more detail.
Erdelyi questions the manner in which repressive coping style is
usually established (by questionnaire measures: individuals with
low self-reported anxiety and high social desirability). He prefers
that individuals be selected by their ability to repress (not-think
of) some target material, regardless of the reason. What would
probably be achieved by such approach is the classification of individuals by their abilities in working memory, executive control
functions, fluid intelligence, and so forth. This would provide a
basis for personality differences in repressive abilities if, but only
if, repression is studied as a general mechanism. However, for
the emotional aspect of repression, the investigation of individual
differences in defensive repression forms a critical link between
experimental and clinical psychology, and this should not be underestimated. Is conventionally measured repressive coping style
indeed associated with a higher capability for general (and not
just defensive) repression, and is the mechanism the same in
BEHAVIORAL AND BRAIN SCIENCES (2006) 29:5
519